Network-modeling veteran at Télécom Paris now publishing on LLM calibration auditing and Wikidata quality (h-index 37)
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
37
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
The output axis would measure career publication volume and productivity, but this snapshot does not score it from the available OpenAlex data; yearly counts exist, yet the derived output dimension is missing, so no score is claimed.
Impact—
The impact axis would capture citation-based influence — how widely and strongly the work is cited. An h-index of 37 and i10-index of 84 appear in the raw record, but the dimension itself is unscored here, so no estimate is made.
Focus0/100
Focus scores 0, the lowest reading in this record, and the raw data explains it: the field distribution splits exactly 50/50 between machine learning (LLM uncertainty and evaluation) and knowledge graphs and semantic web, while the bibliometric topic list is dominated by networking themes — network traffic and congestion control (115) and queuing theory (48). The two reviewed 2026 papers belong to the ML and semantic-web fields rather than the long-run networking topic base.
Collaboration39/100
Collaboration scores 39 on the radar, consistent with a compact recent circle: 18 works over the last three years involve only 6 distinct co-authors. The recurring partners listed are Roman Plaud, Antoine Saillenfest, Matthieu Labeau, Willem Waegeman, Yiwen Peng, and Marc Jeanmougin; on both reviewed 2026 papers he appears in non-corresponding middle or last author positions.
Activity86/100
Activity scores 86, the strongest axis in the record. The OpenAlex countsByYear series runs from 1999 through 2026 with no zero-output year, peaks at 18 works in 2011 and again in 2014, and shows 18 works in the most recent three-year window (2024 alone reached 10), indicating sustained and currently revived productivity.
Reproducibility78/100
Reproducibility scores 78, computed as the mean over 2 Evolve-reviewed papers. The signals are concrete and split: WiCleanData releases open code and is rated high on reproducibility readiness, while the single-query calibration audit is rated medium — its logit_bias threshold method is validated only in simulation, without live-API tests or error bars.
Research summary
Thomas Bonald is a researcher at Télécom Paris whose OpenAlex-indexed record spans 1999–2026, with an h-index of 37 and an i10-index of 84. His bibliometric topic profile is rooted in networking: Network Traffic and Congestion Control (115 topic hits), Advanced Wireless Network Optimization (60), Advanced Queuing Theory Analysis (48), Software-Defined Networks and 5G (41), and Wireless Communication Networks Research (36). The current field distribution, however, splits evenly between machine learning (LLM uncertainty and evaluation) and knowledge graphs and semantic web (0.5 each), matching the two 2026 arXiv preprints Evolve has reviewed: WiCleanData, a taxonomy-refinement pipeline claiming type-violation-free coverage of roughly 420 million Wikidata facts, and Single-Query Black-Box Calibration Auditing via Logit Bias. On both he is a non-corresponding co-author, in last and middle positions respectively, consistent with a senior contributor role. Activity is the strongest radar axis (86): 18 works in the last three years with only 6 distinct co-authors, and annual counts rebounded to 10 in 2024 after the low 2021–2023 stretch. Reproducibility scores 78, the mean across the 2 reviewed papers. Collaboration sits at 39, reflecting the small recent co-author pool, while focus scores 0 given the exact 50/50 field split; the output and impact dimensions remain unscored in this snapshot.
Collaboration pattern
The recent collaboration footprint is compact: 18 works in the last three years involve only 6 distinct co-authors, giving a collaboration radar of 39. Recurring partners include Roman Plaud, Antoine Saillenfest, Matthieu Labeau, Willem Waegeman, Yiwen Peng, and Marc Jeanmougin. On both reviewed 2026 papers he takes a non-corresponding middle or last position, fitting a senior contributor who co-signs rather than leads.
Research keywords
Highlights
WiCleanData: toward a type-consistent 420M-fact Wikidata
One query to audit the calibration of black-box LLMs
Timeline
- 1999Earliest year in the OpenAlex countsByYear series, with 5 recorded works — the start of the indexed publication record.
- 2011Annual output reaches 18 works, the joint highest count in the series.
- 2014A second peak of 18 works, confirming a sustained mid-career run of output.
- 2024Output rebounds to 10 works after the 2021–2023 low (2, 1, and 3 works), lifting the activity radar to 86.
- 2026Two arXiv preprints enter the Evolve catalog: WiCleanData and Single-Query Black-Box Calibration Auditing via Logit Bias.
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometric figures (h-index, yearly counts, topics) come from an OpenAlex identity profile only; no other citation source is included.
- The output and impact radar dimensions are unscored in this snapshot because the underlying data is insufficient; h-index 37 and i10-index 84 are raw values, not dimension scores.
- Field attribution mixes scopes: the topic list is networking-oriented, while the 50/50 field distribution and both reviewed papers sit in ML and semantic web; only 2 Evolve-reviewed papers, both 2026 arXiv preprints, inform the current field split.
- The country code on file is BE while the listed institution is Télécom Paris (France); affiliation metadata was not independently verified.